[phenixbb] B-group and TLS

[Pavel Afonine]

Hi Engin,

1) phenix.refine writes total B-factors into PDB file. If TLS is used, 
then there are ANISOU records that contains total atomic B. The ATOM 
records in this case contain the isotropic equivalent of ANISOU (and NOT 
residual B, as Refmac outputs).

2) When you do group B-factors refinement, the original B-factors are 
not reset. For example, if your starting B-factors BEFORE group B 
refinement are 5,6,7, then after group B refinement they can be 7,8,9  
(if refined group B is 2) or 2,3,4 (if refined group B is "-3").

3) The total B-factor used in phenix.refine:
Utotal = Ulocal + Utls + Ucryst
where Ulocal can be refined individually for each atom or as a group (in 
group B refinement is chosen).

This is explained in details in phenix.refine documentation:
http://www.phenix-online.org/documentation/refinement.htm
see section "Refinement of atomic displacement parameters (commonly 
named as ADP or B-factors)".

For an illustration you can have a look at the slides #28-32:
http://cci.lbl.gov/~afonine/aca2008_knoxville_neutron/

Pavel.



On 12/16/2008 12:08 PM, Engin Ozkan wrote:
> Dear phenixbb people,
>
> I am performing group ADP refinement and TLS at the same time.  At the 
> end, the total B factors in the pdb are not identical within residue 
> groups. 
>
> ATOM      7  N   ILE A  11      23.863   6.894  77.827  1.00151.81      
> A    N
> ANISOU    7  N   ILE A  11    20044  20245  17393   -796  -2448   -568  
> A    N
> ATOM      8  CA  ILE A  11      23.647   6.312  76.502  1.00149.93      
> A    C
> ANISOU    8  CA  ILE A  11    19407  19827  17732   -699  -2577   -541  
> A    C
> ATOM      9  CB  ILE A  11      24.451   7.060  75.423  1.00142.17      
> A    C
> ANISOU    9  CB  ILE A  11    18258  18757  17003   -502  -2834   -817  
> A    C
> ATOM     10  CG2 ILE A  11      24.819   6.135  74.260  1.00140.21      
> A    C
> ANISOU   10  CG2 ILE A  11    17760  18301  17213   -445  -3136   -730  
> A    C
> ATOM     11  CG1 ILE A  11      23.648   8.266  74.929  1.00141.20      
> A    C
> ANISOU   11  CG1 ILE A  11    17949  18698  17002   -369  -2502  -1103  
> A    C
> ATOM     12  CD1 ILE A  11      22.184   7.945  74.630  1.00147.23      
> A    C
> ANISOU   12  CD1 ILE A  11    18421  19495  18025   -392  -2156  -1014  
> A    C
> ATOM     13  C   ILE A  11      24.000   4.831  76.450  1.00152.71      
> A    C
> ANISOU   13  C   ILE A  11    19772  20013  18238   -798  -2860   -201  
> A    C
> ATOM     14  O   ILE A  11      24.895   4.377  77.158  1.00153.29      
> A    O
> ANISOU   14  O   ILE A  11    20114  20074  18055   -857  -3135     -9  
> A    O
>
> I would like to know if the refinement was truly grouped B-factor 
> refinement - I am suspecting the residual B-factors might be grouped, 
> but I haven't checked for that.  Could anyone, maybe Pavel, confirm that 
> is the case?  Is there anyplace in the log files I could see the number 
> of B groups?
>
> Thanks,
>
> Engin
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://www.phenix-online.org/mailman/listinfo/phenixbb
>